These 2 tweets about #COVID, which are essentially identical (& also not really true) were both retweeted 100,000+ times & liked more than half a million times each. One account has ~2k followers; the other ~9k.
And I found both while looking at a Russian language account...
And I found both while looking at a Russian language account...

Holy crap. The two tweets below have engagement rates of 41,427% and 7,052%. To put that in perspective, most tweets have engagement rates of less than 5%, with exceptions for some viral tweets. These are the highest engagement rates I’ve ever seen.
https://twitter.com/rvawonk/status/1260901145247178753?s=21 https://twitter.com/rvawonk/status/1260901145247178753
https://twitter.com/rvawonk/status/1260901145247178753?s=21 https://twitter.com/rvawonk/status/1260901145247178753
I wrote a bit about engagement rates in this article analyzing the (extremely astroturfed) WalkAway Movement.
The highest engagement rate I observed among those accounts was just over 4,000% — until today, that was the highest I’d ever seen. https://arcdigital.media/pro-trump-russian-linked-twitter-accounts-are-posing-as-ex-democrats-in-new-astroturfed-movement-20359c1906d3
The highest engagement rate I observed among those accounts was just over 4,000% — until today, that was the highest I’d ever seen. https://arcdigital.media/pro-trump-russian-linked-twitter-accounts-are-posing-as-ex-democrats-in-new-astroturfed-movement-20359c1906d3
Engagement rates provide a somewhat standardized metric that looks at the total number of engagements with a tweet (e.g., likes/replies/retweets) relative to the total number of followers. This is often a good initial indicator of artificial amplification.
https://arcdigital.media/pro-trump-russian-linked-twitter-accounts-are-posing-as-ex-democrats-in-new-astroturfed-movement-20359c1906d3
https://arcdigital.media/pro-trump-russian-linked-twitter-accounts-are-posing-as-ex-democrats-in-new-astroturfed-movement-20359c1906d3